Culture of Honor and Attitudes Toward Intimate Partner Violence in Latinos

Adherence to culture of honor is examined in relationship to attitudes toward intimate partner violence (IPV) in Latinos. Prior research suggests that high honor culture adherence may be associated with more tolerance for IPV. In the present study, Latino participants (N = 72) completed questionnaires on culture of honor and attitudes toward various aspects of IPV, including attitudes toward the violence itself, toward the perpetrator, toward the victim, and toward victim help-seeking. High and low reputation threat conditions were depicted in vignettes of IPV to manipulate for honor adherence. Findings suggest that participants who adhere more greatly to culture of honor are more accepting of IPV, more approving of the perpetrator, and less approving of seeking help.

Throughout the last three decades, intimate partner violence (IPV) has gained recognition as a major public health concern. According to the 2010 National Intimate Partner and Sexual Violence Survey carried out by the Centers for Disease Control and Prevention, 36% of women and 28% of men have experienced IPV at least once in their lifetime (Black et al., 2011). Aside from the detrimental physical and mental health problems caused by IPV, it has been suggested that IPV also takes a marked financial toll on our families and communities (Centers for Disease Control and Prevention, 2003). It is estimated that injuries caused by IPV account for up to 35% of women's emergency room visits (Shea, Mahoney, & Lacey, 1997), amounting to nearly US$4.1 billion in medical costs each year. Clearly, the physical, mental, and economic costs of IPV warrant the implementation of intervention and prevention programs. Although IPV affects families of all racial and ethnic backgrounds, little research has focused on the prevalence and impact of IPV within minority groups (Hage, 2000;Jasinski & Williams, 1998). As the fastest growing minority group in the United States (U.S. Department of Commerce, U.S. Census Bureau, 2008), research on IPV within Latino groups is especially crucial. Therefore, the following study focuses exclusively on IPV among Latinos, and more specifically, attitudes toward IPV perpetrated by men against women.
Estimates of IPV prevalence within Latino families vary. Although the National Violence Against Women Survey (NVAWS) estimates that about one in four Latina women report victimization (Tjaden & Thoennes, 2000), other sources (Hazen & Soriano, 2007;Ingram, 2007) report higher rates of domestic abuse among Latinas. For example, Hazen and Soriano (2007) found that 33.9% of Latina women report being the victim of domestic violence at least once in their lifetime. A study conducted by Ingram (2007) suggested that as many as half of all Latino men and women have experienced some form of domestic violence at some time in their lives. Because evidence suggests that abused women tend to underreport instances of IPV (Sampselle, 1992), even the most conservative prevalence estimates press the need for immediate intervention and prevention program implementation.
In an effort to create the most effective intervention programs, researchers and health professionals have recognized the increasing need for culturally relevant IPV research (Gillum, 2008). Effective intervention and prevention programs must account for culture-related factors and influences, acknowledging evidence that culture can determine if an instance of IPV is perceived as abusive or not by the victim (Lira, Koss, & Russo, 1999), attitudes toward victims and perpetrators, and attitudes toward help-seeking (Vandello, Cohen, Grandon, & Franiuk, 2009). Therefore, the following study focuses exclusively on IPV among Latinos. By determining whether there are cultural factors associated with attitudes toward IPV, more effective intervention and prevention programs can be designed that acknowledge those cultural factors. Culture of honor, originally used to describe the cultural environment of the southern United States, can be identified in groups that place a great deal of importance on social image in connection to family reputation (Brown, Osterman, & Barnes, 2009;Nisbett, 1993). In honor cultures, displays of aggression or violence can be used to foster a reputation of strength, promoting respect from community members and pride within the family (Cohen & Nisbett, 1994). Violence or aggression may also be used to correct situations that threaten to harm a family's honorable reputation or bring shame upon a family's public image (Cohen & Nisbett, 1994;Henry, 2009;Rodriguez Mosquera, Fischer, Manstead, & Zaalberg, 2008).
Aggression in response to threats to honor has recently been examined in student populations and low socioeconomic status groups. In student samples, honor-adherent participants are more likely to approve of verbally aggressive responses to insults than their non-honor-adherent peers (Henry, 2009;Rodriguez Mosquera et al., 2008). For example, Rodriguez Mosquera et al. (2008) measured honor orientation in a sample of Maroccan/Turkish-Dutch and ethnic Dutch participants with a questionnaire assessing the importance of their family's reputation and importance to be respected and positively evaluated by others. They found that honor-adherent Maroccan/Turkish-Dutch participants were more strongly motivated to punish an insulter in response to a reputation-shaming insult, compared with nonhonor-adherent Dutch participants. Moreover, studies with low socioeconomic status participants suggest that shame or threats to reputation tend to mediate approval of aggressive responses (Henry, 2009). Corroborating earlier literature (Nisbett, 1993), these findings support the concept of a strong relationship between public reputation and self-image in honor cultures. Honor has recently been studied in Latino samples as well, where Latinos' approval of aggression in response to a reputation threat is congruent with honor culture ideals (Vandello & Cohen, 2003).
Honor cultures can also be characterized by strong levels of family cohesion and strong gender roles, as are evident in Latino cultures (Galanti, 2003). The importance of family pride and cohesion, or familismo, have been demonstrated by research on Latino culture (Rivera et al., 2008). Based on solidarity, loyalty, and reciprocity, familismo describes a culturally crucial bond between family members and emphasizes the importance of family image and honor (Cortés, 1995;Galanti, 2003). Although certain aspects of the traditional Latino culture can be captured by the concept of honor culture, this does not imply that this is the only way to describe Latino culture. As argued by Cohen (2009) there are many different cultural dimensions that one might examine, such as religion, individualism-collectivism, or socioeconomic status. And conversely, Latinos are not the only group that share honor culture ideals. For example, as Cohen and Nisbett (1997) demonstrated, Southern and Western regions of the United States can also be characterized by honor culture ideals. Rodriguez Mosquera et al. (2008) demonstrated that individuals of Maroccan/Turkish-Dutch ancestry value honor ideals significantly more than ethnically Dutch individuals.
Despite cross-cultural differences in adherence to honor ideals, having a certain ethnic background such as Maroccan or Turkish or Hispanic does NOT imply universal adherence to a pervasive cultural ideals. As Leung and Cohen (2011) have pointed out, it is important to note that there are individual differences in the degree to which individuals within a culture subscribe and internalize cultural ideals. Therefore, in the present study, we assess these individual differences in adherence to honor culture ideals using a questionnaire developed by López-Zafra (2007). In sum then, honor culture ideals are just one part of the Latino culture, other regions of the world also can be characterized by honor culture ideals and not everybody within a culture subscribe to the cultural ideals.
Another facet of Latino culture revolves around gender roles. Research indicates the existence of strict gender role expectations within Latino culture, suggesting that Latinos place greater importance on gender role adherence in comparison with other cultures (Galanti, 2003;Rivera et al., 2008). In Latino culture, machismo is used to describe the traditional male gender role. While machismo includes positive aspects such as strength, courage, and responsibility, it also includes traits that may be harmful, such as aggression and dominance over women and the family (Torres, Solberg, & Carlstrom, 2002). Latino culture also holds strong expectations for women's gender roles, resulting in a feminine honor code somewhat contrary to the male honor code. Inspired by the Virgin Mary, marianismo refers to cultural tradition dictating that women should be pure and selfsacrificing, devoting their lives to the family (Galanti, 2003). While the feminine honor code promotes loyalty, humility, and fidelity, its emphasis on deference to men can also encourage unconditional compliance, unassertiveness, and submissiveness.
The differentiation between machismo and marianismo parallels the distinction drawn between masculine and feminine honor codes in honor cultures by Rodriguez Mosquera, Manstead, and Fischer (2002). Rodriguez Mosquera et al. (2002) described the masculine honor code as centering on virility, authority over, and protection of family while the feminine honor code is described as centering on modesty and sexual shame. Violation of gendered honor codes, particularly the feminine honor code (e.g., sexual immodesty), can damage self-image and threaten family reputation (Rodriguez Mosquera et al., 2002).
Gendered honor codes can influence Latinos' attitudes toward family honor and, in turn, IPV. Many of today's Latina women adhere strongly to certain aspects of the traditional Latina women's gender role, including a strong sense of dedication to the family (Bauer, Rodriguez, Quiroga, & Flores-Ortiz, 2000). Other research suggests that some Latina women adhere strongly to more traditional marianismo ideals, believing a woman should submit to her husband's authority, stand behind his decisions, and tolerate his actions (Galanti, 2003;Moreno, 2007). The expectation for women to accept their husband's behavior can be problematic, particularly when evidence suggests that cultures stressing traditional women's gender roles are associated with high rates of violence (Ahrens, Rios-Mandel, Isas, & Lopez, 2010;Cohen, 1998). The combination of strong gender role expectations and a focus on familism lead us to assert that Latino culture holds honor at a high level of importance. Consistent with the familismo script, honor is of particular concern when Latinas seek help for IPV; seeking help from just any agency, organization, or authority could have detrimental effects on the family's reputation within the community. Thus, Latinos are more likely to seek help for IPV from a family member and less likely to seek help from formal agencies than are non-Latinos (Ingram, 2007).
In Latinos, attitudes toward IPV differ with abuse type (Vandello et al., 2009). In general, Latinos who adhere to honor traditions are more accepting of honor-related IPV than non-honor-related IPV. Honor-related IPV can be understood as abuse that is jealousy-related or spurred by an offense that compromises a man or his family's honor. If, for instance, a married woman flirts with another man, her husband may respond abusively toward her to reestablish a sense of honorable control over his family. As stated previously, using violence to restore a reputation can be viewed as an acceptable practice in honor cultures, which may explain why some Latinos condone IPV in honor-related scenarios (Vandello & Cohen, 2003).
Evidence indicates that those accustomed to honor cultures (including Latinos) tend to view abusers more favorably when the abuser's violent actions are in response to an affront to honor or a threat to the family's reputation (Vandello et al., 2009). Vandello et al. (2009) indicated that participants from honor cultures consistently rate abusive husbands more positively when their abuse is honor-or jealousy-related. It is important to note that these patterns only pertain to honor-or jealousy-related abuse where a family's public image is threatened; abuse unrelated to honor or reputation is not typically condoned by honoradherent participants (Vandello et al., 2009). The same study finds that participants from honor cultures are less approving of help-seeking in honor-related scenarios; honor-adherent participants rate a wife more positively when she stays with her abusive husband than when she leaves him (Vandello et al., 2009).
As suggested by Vandello et al. (2009), honor adherence may influence individual attitudes toward IPV. In addition, research with honor-adherent groups highlights the importance of reputation and respect within the community (Henry 2009;Rodriguez Mosquera et al., 2008). Thus, the purpose of the present study is to examine how Latino participants' adherence to culture of honor ideals influences their attitudes toward IPV. As discussed above, Leung and Cohen (2011) suggested that the degree to which members of a cultural group subscribe to cultural ideals can vary greatly. Therefore, in the present study, we assess this individual variability in endorsing the culture of honor ideals with a scale developed by López-Zafra (2007). Thus, we will determine influences of culture of honor ideals on acceptability of IPV, perceptions of the perpetrator, the victim, and victim help-seeking when reputation is or is not threatened.
To determine the impact of culture of honor ideals, we created a manipulation of perceived damage to honor within a reputation-threatening scenario. In the high reputation threat condition, participants read about an IPV incident that takes place in reaction to a public marital transgression by a wife, that is, kissing another man in public. In the low reputation threat condition, the infidelity occurs in private and therefore has no potential for damaging the reputation of the family in the community. Consistent with Vandello et al. (2009), Rodriguez Mosquera et al. (2008, and Henry (2009), a threat to reputation can indicate damage to honor, whereas an unthreatened reputation does not damage honor. In addition, we examined how respondents' adherence to honor culture ideals (high or low) influences their perception of damage to honor and IPV attitudes in the high and low reputation threat conditions. Perception of honor damage functions as the manipulation check for the high and low reputation threat conditions. We expect that perceived honor damage and general IPV acceptance will be greater in the high reputation threat condition than in the low reputation threat condition. As suggested by Vandello et al. (2009), we further expect that participants with high culture of honor adherence will express greater IPV acceptance, higher perpetrator approval, and lower victim and help-seeking approval compared with participants with low culture of honor adherence. Thus, we expect that participants with high culture of honor adherence in the high reputation threat condition will perceive greater honor damage, greater IPV acceptance, and express higher perpetrator approval and lower victim and help-seeking approval compared with participants in the low reputation threat condition. Finally, given the findings of strict gender role expectations within Latino culture (Galanti, 2003), we will explore gender differences among our Latino participants regarding honor damage, IPV acceptance, victim and perpetrator approval, and help-seeking approval in the high and low reputation threat scenarios.

Participants
Participants were recruited through three organizations located in St. Paul, Minnesota, chosen specifically for their work with and services for the Latino population. Overall, 72 adults participated in the study (45 women and 27 men). To ensure an accurate data set, data belonging to 20 participants who did not closely follow participation instructions (including data from participants who left several items blank or unfinished) were eliminated from the data analysis. Participant age ranged from 18 to 58, with an average age of 31 years (SD = 8.43). While all participants reported belonging to the Hispanic/Latino ethnic group, some also indicated belonging to the White (including Middle Eastern) and Asian ethnic groups (1 Hispanic/Latino and Asian, 1 Hispanic/ Latino and White). Approximately 45% of participants were born in Mexico, 17% were born in Ecuador, and 17% were born in the United States. Of the remaining 21% of participants, 14% were born in El Salvador, Guatemala, Nicaragua, or Honduras, and 7% were born in Bolivia, Columbia, Puerto Rico, or Venezuela. In accordance with ethical guidelines established by the American Psychological Association, all participants read and signed informed consent forms prior to participation in the study. In addition, participants were offered a US$5 gift card to a major retail store to reward participation.

Design
We utilized a 2 × 2 between-subjects factorial design. The subject variable, honor adherence, had two levels: high and low adherence to honor. These levels were determined via a median split procedure on scores of an honor adherence measure (explained below). We crossed this subject variable with the two levels of the independent variable, reputation threat, manipulated with the use of IPV vignettes (explained below). The participants were randomly assigned to read one of two kinds of IPV vignettes-one vignette describing an IPV incident involving a small threat to the man's reputation and the other vignette involving a substantial threat to the man's reputation. Participants were randomly assigned to read just one of the vignettes to minimize awareness of the purpose of the study.

Variables and Measures
Demographics and Culture of Honor Adherence. A 6-item demographics questionnaire regarding participant age, gender, place of birth, and ethnic group(s) was used. A 17-item scale (López-Zafra, 2007) developed specifically for Hispanics was used to determine participants' adherence to honor-related attitudes. The honor scale measures participants' agreement with various honor-related statements on a 1 to 5 scale, with 1 indicating disagreement and 5 indicating agreement using items such as "If someone insults my family or me, I will confront him/her"; "In general, honor is of high importance"; and "An insult to honor should be strongly punished by the group." Computation of the internal reliability of this 17-item scale showed that this scale had high internal reliability with α = .81.

Vignettes Used to Manipulate High Versus Low Reputation
Threat. To determine the effect of a reputation threat, vignettes of an IPV incident were developed for both the high and low reputation threat conditions. As suggested by Rodriguez Mosquera et al. (2002), public violation of the feminine honor code of sexual shame are especially likely to threaten the core values of the code of honor. Therefore, in the high reputation threat condition, we describe an incident of IPV in which a husband, Mario, acts abusively toward his wife, Ana, in response to discovering evidence about the public nature of her extramarital affair. In this condition, participants are told that the entire town knows about the affair, posing a strong threat to the family's honor. In the low reputation threat condition, the identical IPV incident takes place, but in response to a secret affair. In this condition, participants are assured that no one else knows about Ana's infidelity. Thus, the affair is not likely to cause damage to the family's honor. Previous research (Vandello & Cohen, 2003) has emphasized the importance of women's role in preserving the family's honor by keeping the feminine honor code; thus, Ana's infidelity is a salient violation of honor to both genders. Please see the appendix for the vignettes.
Damage to Honor. To check the effectiveness of the high versus low reputation threat manipulation, we developed a manipulation check index, Honor Damage, which was composed of items regarding participants' perception of the damage to the family's reputation, including "Ana has violated her family's honor"; "Ana has violated Mario's honor"; and "Ana has damaged her family's reputation." Again, a 5-point response scale was used with 1 indicating disagreement and 5 indicating agreement. The internal reliability computation showed high internal reliability for this threeitem index (α = .87).
Finally, a general IPV acceptance questionnaire was adapted from Vandello et al. (2009). The questionnaire consists of several statements regarding attitudes toward the victim, perpetrator, IPV, and help-seeking behavior. The statements aim to measure these attitudes using a 1 to 5 scale, with 1 indicating disagreement and 5 indicating agreement. Using various items, we created several indexes to capture various facets of IPV attitudes.
IPV Approval. The IPV Approval index contained three items pertaining to the acceptability of the IPV itself: "Mario was right in showing Ana who's the boss in the family"; "As long as hitting Ana didn't cause permanent damage, it was okay that Mario hit Ana"; and "Ana's behavior was an acceptable reason for Mario to hit her." Internal reliability of this IPV Approval index was high with α = .80.
Perpetrator and Victim Perception. Perpetrator Perception was composed of four items pertaining to the perpetrator's goodness, including "Mario is trustworthy"; "Mario is responsible"; "Mario is reasonable"; and "Mario is good" (α = .75). Victim Perception was composed of six items concerning the victim's goodness: "Ana is smart"; "Ana is competent"; "Ana is strong"; "Ana is good"; "Ana is reasonable"; and "Ana is submissive" (α = .85).
Approval of Seeking Help. An index composed of six items assesses approval of the victim seeking help from clergy, including "Ana is trustworthy/selfish/reasonable/smart for seeking help from a priest/minister/clergy person"; "It is acceptable for Ana to seek help from a priest/minister/clergy person"; and "Seeking help from a priest/minister/clergy person would damage Ana's family's reputation" (α = .69). In addition, approval of help-seeking from a shelter was assessed with a six item index (α = .76) similar to the Clergy Help index, identical with the exception of the help-seeking location listed in each statement. For example, the first item included was "Ana is selfish for seeking help from a battered women's shelter or helpline." Finally, approval of seeking help from a sister was assessed (α = .79) with the same six items contained in the Clergy Help index, identical with the exception of the help-seeking location listed in each statement.
To ensure accurate translation, any materials not available in Spanish were translated using the back-translation method. We obtained the Honor Adherence scale in Spanish from the author (López-Zafra, 2007) while the remaining materials (i.e., the consent form, the vignettes, the damage to honor, IPV approval, perpetrator and victim perception, and help-seeking approval measures) were subjected to back-translation. First, we enlisted a native Spanish speaker fluent in English to translate the materials from English into Spanish. Subsequently, another bilingual individual translated the materials from Spanish into English. We then checked for correspondence in meaning of the backtranslated materials against the original English version. Any discrepancies were resolved through a collaborative meeting in which the translators came to a consensus on the best Spanish phrasing to use.

Procedure
With permission from each organization's director, tables were set up in lobbies or waiting areas with bilingual signage inviting participants. Participants who approached the table were initially greeted in Spanish, as Spanish was the main language spoken at each location. If a participant responded to the initial greeting in English, the invitation to participate was continued in English. Otherwise, the invitation to participate continued in Spanish. Participants who agreed to participate in the study were offered a choice of completing the consent form in either English or Spanish. Participants were then given the questionnaire packet (in the same language as the consent form they chose) to complete.
The questionnaire measuring demographics, honor, and IPV acceptance was attached to either a high or low reputation threat scenario to create two separate types of questionnaire packets. The two questionnaire conditions, high and low reputation threat, were placed in random order prior to arriving at each location for data collection. Thus, participants were randomly assigned to either the high or low reputation threat manipulation. At the conclusion of participation, each participant was offered a US$5 gift card to a major discount retailer in addition to a slip of paper indicating (in the same language as their questionnaire) the appropriate people or organizations to contact should they have questions regarding the study or IPV.

Results
The 17 items composing the Honor Adherence scale (López-Zafra, 2007) were summed and averaged, with mean scores ranging from 1.88 to 4.35. The mean Honor Adherence score was 3.20 (SD = 0.61) with high internal reliability (α = .81). For the purpose of participant categorization, a median split was performed using participants' Honor Adherence scores. Participants scoring higher than 3.14 (the 50th percentile of the score distribution on the Honor Adherence scale) were categorized as high honor culture adherent while participants scoring lower than 3.14 were categorized as low honor culture adherent.

Examining the Effects of Culture of Honor Adherence and Reputation Threat
Manipulation Check: Honor Damage. As described above, the Honor Damage index constituted our manipulation check of the effectiveness of the high versus low reputation threat manipulation. The overall mean of this manipulation check was 3.23 (SD = 1.50). Scores ranged from 1 to 5, with 1 indicating a low rating of damage to honor and 5 indicating a high rating of damage to honor. The two-way ANOVA showed no significant main effects for high versus low reputation threat, and no significant main effects for low versus high culture of honor adherence. More importantly, though, the interaction between the reputation threat manipulation and culture of honor adherence was nearly significant, with F(1, 70) = 3.72, p = .058. The trends in the means suggest that participants high in adherence to culture of honor ideals perceived more damage to the family's honor than participants low in culture of honor adherence in the high reputation threat condition. In the low reputation threat condition, participants' degree of adherence to culture of honor did not influence perceived honor damage. Please see Figure 1 for the cell means.
IPV Approval. The mean score for the IPV Approval index was 1.72 (SD = 1.16). Scores ranged from 1 to 5, with 1 indicating low levels of IPV approval and 5 indicating high levels of IPV approval. The ANOVA showed no significant main effect for the type of reputation threat manipulation on IPV approval. However, as predicted, the significant main effect for honor culture adherence indicated that participants categorized as high in honor adherence were more approving of the IPV (M = 2.05, SD = 1.27) than participants with low honor adherence (M = 1.41, SD = 0.97), with F(1, 69) = 6.29, p = .02. Although the interaction between the categorization of honor culture adherence and the manipulation of reputation threat was not significant, F(1, 69) = 0.81, p = 0.37, the trends in the cell means suggest that highly honor-adherent participants appear to be more approving of IPV than participants low in honor adherence in the high reputation threat condition. In the low reputation threat condition, however, the difference in IPV approval between the low and high honor-adherent groups is much smaller. Please see Figure 2 for the cell means.
Perpetrator Perception. The mean score on the Perpetrator Perception index was 2.60 (SD = 1.15). Scores ranged from 1 to 5, with 1 indicating a low approval rating of the perpetrator and 5 indicating a high approval rating of the perpetrator. The ANOVA showed no significant main effect for the manipulation of reputation threat. However, we did obtain a significant main effect for low versus high honor culture adherence, with F(1, 70) = 4.75, p = .03. Participants categorized as high in honor culture adherence rate the perpetrator more positively (M = 2.90, SD = 1.16) than participants low in honor culture adherence (M = 2.33, SD = 1.09). The interaction between honor culture adherence and type of reputation threat manipulation was not significant.
Victim Perception. The mean score for the Victim Perception index was 2.20 (SD = 0.88). Scores ranged from 1 to 4.67, with 1 indicating a low level of victim approval and 4.67 indicating a high level of victim approval. The ANOVA showed no significant main effect for manipulation of reputation threat on the Victim Perception index, no significant main effect for culture of honor adherence, and no significant interaction between culture of honor adherence and type of reputation threat manipulation. Overall, participants' rather low approval of the victim regardless of condition appears to be a basement effect for this measure.
Approval of Help-Seeking. On all three indexes, scores ranged from 1 to 5, with 1 indicating a low level of approval to seek help and 5 indicating a high level of approval. The mean score for the Clergy Help index was 4.06 (SD = 1.02). The significant main effect for the manipulation of the type of reputation threat, F(1, 69) = 7.19, p = 0.01, showed that participants were more approving of seeking help from a clergy member in the high reputation threat condition (M = 4.40, SD = 0.83) than in the low reputation threat condition (M = 3.74, SD = 1.09). Neither the main effect for low or high adherence to culture of honor nor the interaction was statistically significant. Regarding approval of seeking help from a shelter, the mean score for the Shelter Help index was 3.72  (SD = 1.20) The marginally significant main effect for high and low reputation threat showed that participants appeared to be more approving of seeking help from a shelter in the high reputation threat condition (M = 4.02, SD = 1.13) than in the low reputation threat condition (M = 3.44, SD = 1.22), with F(1, 66) = 3.64, p = .06. As on the Clergy Help index above, the main effect for low or high adherence to culture of honor and the interaction were not statistically significant. On the final approval of help-seeking index, the mean score for the Sister Help index was 3.89 (SD = 1.08). The ANOVA showed no significant main effects or interaction.

Correlational Analyses
As described above, we categorized participants into low or high honor adherence groups based on a median split procedure for the purpose of carrying out an analysis of variance. However, categorizing participants into low or high adherence brings with it some loss of specificity of data. Therefore, we carried out several correlational analyses, keeping participants' precise culture of honor adherence scores intact.
To examine whether the degree to which participants adhere to culture of honor ideals was related to any of the indexes described above, Pearson correlation coefficients were computed using one-tailed tests. As expected, the positive relationship between Honor Adherence and IPV Approval was significant, with r(68) = .20, p = .05, such that the greater the participant's honor adherence, the greater the acceptance of IPV. In addition, a significant, positive correlation between Honor Adherence and Perpetrator Perception was found with r(69) = .32, p = .003, such that the greater the participant's honor adherence, the more positive the perception of the perpetrator. Statistically significant, negative correlations also were found between Honor Adherence and all help-seeking indexes. The correlation to Clergy Help was significant, with r(68) = −.19, p = .05, such that the greater the participant's honor adherence, the lower the acceptability of seeking help from a priest, minister, or clergy person. The negative relationship between Honor Adherence and Sister Help was significant, with r(68) = −.21, p = .04, such that the greater the participant's honor adherence, the lower the acceptability of seeking help from a sister. The negative relationship between Honor Adherence and Shelter Help was also significant, with r(65) = −.22, p = .04, such that the greater the participant's honor adherence, the lower the acceptability of seeking help from a shelter or helpline. Overall then, adherence to culture of honor ideals was associated with disapproval of help-seeking from any sources outside of the immediate relationship.
Next, we examined correlations with the manipulation check index of perceived damage to honor. A significant positive correlation between Honor Damage and IPV Approval was found with r(68) = .24, p = .02, such that the more damaged the participants perceived the family's reputation, the greater the approval of IPV. In addition, a significant positive relationship between Honor Adherence and Perpetrator Perception was found, with r(69) = .41, p = .001, such that the more damaged the participants perceived the family's reputation, the more positive the perception of the perpetrator. Finally, a significant negative relationship between Honor Damage and Victim Perception was found, with r(65) = −.28, p = .01, such that the more damaged the participants perceived the family's reputation, the more negative the perception of the victim.

Independent-Samples t Tests on Gender of Participants
Research suggests that gender may influence perception of honor damage as well as attitudes toward aggression (Vandello, Cohen, & Ransom, 2008). Although we did not predict any effect of gender, we conducted an exploratory analysis of gender difference. Because the low number of male participants (n = 27) prevented us from subjecting the data to a three-way ANOVA (Gender × Honor Adherence × Reputation Threat), independent-samples t tests were conducted to determine the effect of gender on the various indexes. No significant differences were found on the IPV Approval, Perpetrator Perception, Victim Perception, or the help-seeking indexes. A significant difference was found on the manipulation check index of perception of honor damage, with t(69) = 2.02, p = .05. Men perceived more damage to honor (M = 3.68, SD = 1.44) than did women (M = 2.95, SD = 1.49).

Discussion
We expected that perception of honor damage would be influenced by the manipulation (high and low reputation threat) as well as participants' honor adherence level (low or high adherence to culture of honor ideals). Specifically, we predicted that high honor adherence would be associated with greater perceived honor damage in the high rather than low reputation threat condition. Although not statistically significant at .05, the means were in the expected direction, with participants high in adherence to honor culture ideals perceiving more damage to the family's honor in the high reputation threat condition than in the low reputation threat condition. Congruent with previous research suggesting that honor cultures place a considerable degree of concern on social image (Rodriguez Mosquera et al., 2008), our findings highlight the influence of public reputation threat on personal honor.
In the present study, we found the reputation threat manipulation to have no influence on perception of honor damage for participants categorized as low in adherence to culture of honor ideals. That is, participants low in honor adherence perceived moderate amounts of damage to the family's honor regardless of the reputation threat manipulation. However, among participants high in honor adherence, our manipulation of reputation threat appears to have led to greater perceived damage in response to the high reputation threat vignette describing the public nature of the extramarital affair compared with the low reputation threat condition describing the extramarital affair as private. These findings are clearly consistent with previous research indicating that reputation threat is a more salient factor for individuals high in honor adherence than those low in honor adherence (Rodriguez Mosquera et al., 2008;Vandello et al., 2009). These results also suggest a sense of connection between public and private image in honor cultures, that is, a person's self-image is highly influenced by his or her social image, as suggested by Rodriguez Mosquera et al. (2002).
We also predicted a relationship between honor adherence and various measures of general IPV acceptance, including attitudes toward help-seeking and attitudes toward the victim and perpetrator. We expected that participants high in honor adherence would show greater IPV and perpetrator approval as well as lower victim and help-seeking approval. As predicted, honor adherence was positively related to IPV Approval and Perpetrator Perception. That is, individuals high in honor adherence are more condoning of IPV and the violent perpetrator regardless of whether the violence occurs in response to a scenario posing a high or low threat to the family's reputation. These results support previous research suggesting that IPV may be used to "correct" an affront to honor (Vandello & Cohen, 2003). Participants categorized as high in honor culture adherence may have considered the violence a way to "right" the infidelity and perhaps even prevent the continuation or escalation of the infidelity, simultaneously lifting blame from the perpetrator and justifying the use of IPV. While these results may be congruent with honor culture concepts, our findings are contrary to recent research that suggests honor-adherent individuals are more condoning of violence in only honor-threatening scenarios (Vandello et al., 2009). To more fully understand IPV as a means to "correct" an affront to honor, further research on the use of violence and aggression as corrective or preventive tools in intimate relationships is needed.
As predicted, adherence to honor was negatively related to help-seeking approval. This finding supports the concept that honor-adherent individuals are less approving of victim help-seeking. These results are congruent with previous research suggesting that honor-adherent individuals may be particularly concerned with family reputation, which can be threatened by victim help-seeking (Ingram, 2007). Ingram (2007) indicated that seeking help for family issues such as IPV, particularly from "outside" sources (e.g., health clinics, battered women's shelters), can bring unfavorable attention or shame to the family's reputation. Considering the damage already implied by the reputation threat manipulation, it makes sense that participants high in honor adherence seem to be wary and disapproving of help-seeking in general.
With this in mind, we further predicted that participants would be more approving of help-seeking in the low reputation threat condition versus the high reputation condition.
Unexpectedly, participants were actually more approving of help-seeking in the high reputation threat condition, particularly for the Clergy Help measure. Participants' support of seeking help from a clergy member is consistent with previous research, suggesting that Latinos consider a priest, minister, or clergy person to be an "inside" source of support (Ingram, 2007). Respondents may have been more supportive of help-seeking from a clergy person as this resource provides help that is twofold; the victim has the opportunity to seek help and counsel for the IPV as well as the chance to repent for her infidelity.
In line with the strong gender role expectations found within Latino culture (Galanti, 2003;Moreno, 2007;Rivera et al., 2008), we found gender effects for the Honor Damage index. According to the results obtained with the t test, men perceived more damage to the family's honor than did women. Explanations for these gender effects may be related to participants' gendered sense of empathy for the female victim or the male perpetrator (Galanti, 2003). Research indicates that women tend to show more empathy for victims than do men (Smith & Frieze, 2003), suggesting that women respondents may have perceived less damage to the family's honor due to their sympathy for the woman victim. This finding is consistent with traditional women's gender roles stressing kindness and empathy (Bem, 1974;Galanti, 2003). Future studies may wish to examine participants' emotional reactions to the scenarios (e.g., "I am angry at Ana") with particular regard for gendered emotional reactions, as suggested by Rodriguez Mosquera et al. (2002). Such research may help develop a more nuanced understanding of gendered emotional reactions to honor damage. In addition, the low number of male participants prevented us from carrying out a fully crossed three-way ANOVA. In future IPV research, a full examination of gender effects in light of culture of honor and reputation threat is warranted.
One limitation of our research relates to the wide assortment of participant education levels and ability to comprehend our measures. For example, several participants spent excessive time completing the questionnaire. This may have been influenced by participants' reading comprehension levels. In addition, some participants had difficulty understanding how to use the rating scales present in our questionnaires. These complications may have led some participants to leave many items incomplete, resulting in the omission of their responses from the data set. Future studies may consider tailoring materials (including all instructions and consent forms) to the average education level of their intended participants. Additional limitations of the present study are that other important variables that may influence attitudes toward IPV were not examined. For example, research suggests that acculturation (Peek-Asa, Garcia, McArthur, & Castro, 2002), recency of immigration status (Littleton, Breitkopf, & Berenson, 2007), socioeconomic status (West, Kaufman Kantor, & Jasinski, 1998), urban versus rural background (Moracco, Hilton, Hodges, & Frasier, 2005), and age (Lown & Vega, 2001) may all be associated with attitudes toward IPV and therefore should be accounted for in future research.
Overall, our study suggests that individuals high in honor adherence are more accepting of IPV and more approving of perpetrators in both high and low reputation threat scenarios. One possible explanation for the lack of a statistically significant effect of the reputation threat manipulation is that any kind of a breach of fidelity, be it a high reputation threat infidelity or a low reputation threat infidelity, violates the norms of the culture of honor. Therefore, the distinction in the reputation threat may not have differentially influenced high culture of honor participants' response. Similarly, among participants with low culture of honor adherence, the distinction between low or high reputation threat infidelity may not have differentially affected participants' responses because no kind of infidelity justified IPV in those participants' minds.
Furthermore, the present study suggests that help-seeking approval is influenced by both honor adherence and reputation threat, with individuals high in honor adherence more approving of help-seeking in scenarios threatening to family reputation. Although our study represents important findings on the connection between honor and IPV acceptance, further research is warranted to continue examining the importance of reputation and gender roles in relationship to honor. Continued research with Latinos in the areas of social reputation, gender roles, and honor will lead to a more nuanced understanding of Latino culture. Moreover, a thorough grasp of culture's role in IPV acceptance will aid in the creation of more effective intervention and prevention programs. Our findings are also congruent with Sokoloff and Dupont's (2005) suggestions that service providers such as counselors should become aware of the influence of cultural differences on the therapeutic process and that domestic violence support requires culturally competent services. It is important to acknowledge then that individuals' adherence to culture of honor principles may constitute an obstacle to help-seeking. Therefore, it is important for intervention and prevention programs to identify the degree to which their constituents adhere to culture of honor ideals to identify those that may be more reluctant to seek help. More specifically, in a recent focus-group study with Hispanic females by Gonzalez-Guarda, Vasquez, Urrutia, Villarruel, and Peragallo (2011), it was determined that rompiendo el silencio (breaking the silence) was an important component in breaking the cycle of abuse. These authors also pointed to the importance of awareness of cultural factors such as machismo in the design of abuse intervention and prevention programs. Another study by Burke, Oomen-Early, and Rager (2009) also identified the notion of silence as "one of the strongest cultural-related themes" emerging from interviews with Latino survivors of abuse. In sum, our findings in conjunction with research by Gonzalez-Guarda et al. (2011) and Burke et al. (2009) clearly pointed toward the importance of cultural factors in help-seeking which in turn is an important, sometimes life-saving, response to abuse.